The library[1] is designed for the same purpose of flink-cep [2]. They both work as single stream operator. Fink-cep defined a NFA model to support pattern match, while flink-siddhi provides much richer features.
To make it simple, here only supports to define CEP operator with Flink API in generic streaming application without dynamical metadata management. [1] https://github.com/apache/bahir-flink/tree/master/flink-library-siddhi [2] https://github.com/apache/flink/tree/master/flink-libraries/flink-cep -- Hao On Tue, Nov 28, 2017 at 4:43 AM, Edward Zhang <[email protected]> wrote: > Thanks for the update, Hao. > > It is an interesting work, but I have a question. Flink is supposed to have > a lot of advanced APIs which supports windowing, SQL-like functions like > Siddhi. So what is the difference integrating Siddhi with Flink would make? > Eagle uses Siddhi to calculate window based event stats for monitoring and > data flow can be dynamically modified based on schema, but this integration > between Siddhi and Flink only have static data model/schema? > > Thanks > Edward > > On Sun, Nov 26, 2017 at 7:00 PM, Hao Chen <[email protected]> wrote: > > > Hi, > > > > Inspired by eagle's streaming alert engine with siddhi, I created a > generic > > light-weight siddhi ondistributed streaming library (currently support > > merged into flink community): https://github.com/apache/ > > bahir-flink/pull/22, > > especially for use case that would like to adopt siddhi in generic > > streaming processing application but not only monitoring and alerting. > > > > - Hao > > >
